This project endeavors to understand the interactions between cytotoxic T lymphocytes (CTL) and the central nervous system (CNS). The first aim is to understand the mechanism(s) used by CTL in the provision of immunity within the CNS. CTL which are blocked in the translation of tumor necrosis factors alpha (TNF-alpha) and TNF-beta will be derived using a retrovirus vector expressing anti-sense TNF. Similarly, antigen- specific CD8+ T cells will be derived from perforin knockout mice. These T cells will be examined for their ability to recognize cells within the CNS and modify an acute viral-induced demyelination. Similarly, the susceptibility of individual class I expressing cell types within the CNS will be tested for recognition by CTL using transgenic technology to express antigen under control of the GFAP promoter exclusively in astrocytes. These mice will be compared to mice expressing the identical antigen in microglia. These data will demonstrate whether these cell types can serve as targets within the CNS and provide evidence for the participation of adhesion molecules in the access and possible regional distribution of lesions within the CNS and if a CTL response restricted to one cell type can cause demyelination by effecting a separate cell type, the oligodendroglia. The sencond aim examines the hypothesis that antigens localized to the parenchyma induce and expand CTL within the local environment in contrast to antigens which induce meningeal infiltrates derived from the cervical lymph nodes (CLN). To examine this hypothesis we will test the localization of immunodominant epitopes from two haplotypes to the parenchyma vs the meanings and determine if the quality of antigen influences localization using a novel defective interfering vector system.